Comment It's going to be a long time before Oracle can take on the likes of HP and IBM for the IT-market crown. But before he retires, you can bet Larry Ellison's last billion bucks that he most surely wants to become the dominant systems supplier in the data center.
This goal is not as unattainable as it might seem, especially now that Ellison, Oracle's co-founder, chief executive officer, and largest shareholder, has hired his tennis buddy and former top dog at HP, Mark Hurd, to be co-president of the software giant and hardware wannabe. Hurd was announced as the peer to Safra Catz and the replacement for Charles Philips, who resigned on the Labor Day holiday.
The Wall Street Journal reported Monday that Phillips did indeed want to leave Oracle and was not pushed out of the company. Phillips had apparently been chasing the CEO job at CA Technologies (formerly known as Computer Associates and, like Oracle, a database pioneer from the mainframe days), but a many-year affair that Phillips had while he was married was exposed by his mistress, who alleged that Phillips claimed to be divorced.
Phillips is a former captain in the US Marines, a graduate of the US Air Force Academy with a BS in compsci, and managing director at investment bank Morgan Stanley, where he was a software analyst. He seemed particularly well-suited to the Oracle culture — and in fact, you could say he helped create its most recent incarnation. But perhaps Phillips chafed at not being in charge himself, and when Hurd opened the door to try to come into Oracle, Phillips waltzed out, joining the ranks of former Oracle execs who had had quite enough of Larry Ellison.
Which Hurd might also do himself years hence, having already run a $125bn IT behemoth that is strong everywhere Oracle is not, and that is nearly five times as large as Oracle in terms of revenues. So you have to figure that the plan to bring Hurd onboard is a very long-term one, perhaps with Hurd taking some work off Ellison's plate so he can devote more time to boats, planes, playing guitar, and philanthropy. And just maybe Ellison and Hurd — perhaps with Catz, and perhaps not — have cooked up a scheme to get Oracle up to the $100bn mark sometime before Ellison retires so he can leave a really big endowment to charity alongside Warren Buffett and Bill Gates.
People forget that Hurd has seen a lot of the IT industry, and while not a peer of Ellison's, he isn't a banker or software analyst either, as Catz and Phillips were before joining Oracle. Hurd has participated in the IT industry since he left Baylor University with a BA in business administration, only three years after Larry Ellison and Bob Miner cofounded Oracle.
Everybody these days always forgets Miner, who basically coded Oracle's first two iterations of the eponymous relational database all by his lonesome. Like Hurd, Ellison got his start in sales and thinks he is a techie. Perhaps after more than 30 years and nearly a billion in the bank for each of those years, Ellison is.
Hurd was hired by NCR back in 1980, and was a salesman when that company became an open-systems pioneer of sorts with the NCR Tower "supermicrocomputer system," as the company called it, which ran Unix. NCR sold some 100,000 units of the first machine, the Tower16/32, and that gave the Cash a beachhead in the financial-services industry — one that Sun Microsystems took over a decade later when it moved from Unix workstations into Unix servers.
Hurd was around in 1991 when AT&T snapped up NCR for a stunning (at the time) $7.4bn and shut down its own computer-server business. He was also there when NCR, still being eaten by Ma Bell (or the remnants of her), ate data-warehousing specialist Teradata, eventually moving the company's clustered database to NCR iron. And not only was Hurd there to run Teradata for three years after Ma Bell spat NCR back out like an allergic reaction in 1997, but he was also good enough at it to rise to the top at NCR, being named president in 2001, chief operating officer in 2002, and chief executive officer in 2003.
NCR spun out Teradata two years after Hurd left to become president and CEO at HP in 2005. This was perhaps a very stupid thing to do, but it made some shareholders and investment bankers some quick cash — just like Oracle buying up Teradata might do to give the software giant some real credibility in data warehousing.
Who better to manage a Teradata acquisition than Hurd? No one else on earth.
"I believe Oracle's strategy of combining software with hardware will enable Oracle to beat IBM in both enterprise servers and storage," said Hurd in the statement Oracle put out announcing his appointment. "Exadata is just the beginning. We have some exciting new systems we are going to announce later this month at Oracle OpenWorld. I'm excited to be a part of the most innovative technology team in the IT industry."
Why would Oracle need Teradata when it already has the Exadata V2 appliance, and is probably cooking up the Exadata V3 appliance (based on the forthcoming "Rainbow Falls" Sparc T3 processors running Solaris and some Xeon-based blade running Oracle Enterprise Linux, if I had to guess) for announcement at Oracle OpenWorld in two weeks? The reason, as all of the data-warehousing players have said repeatedly since Exadata V1 was launched on HP iron (with Ellison and Hurd being best buddies back then), is that the Exadata software is not great at data warehousing even if it is great at scaling and protecting OLTP systems. And Teradata has a customer base and a revenue stream, just like Sun did.
If Oracle learned anything from glomming together PeopleSoft, Siebel Systems, and JD Edwards, it was to continue to develop these products so as not to upset customers and to keep them paying their maintenance and doing their upgrades. But over time, Oracle nudges them all towards a common product that makes use of all the smarts in all the products. That's the plan with the Fusion applications and middleware, and it will be the plan with the Fusion systems, which Oracle will no doubt eventually call whatever machines it makes years hence.
But bringing Hurd on board is about way more than acquiring a company like Teradata to bolster the data-warehousing business. (I say "like Teradata" since the company has a market capitalization of $5.8bn as we go to press against annual revenues of $1.71bn and net income of a mere $254m in 2009.) First and foremost, Hurd is perfectly capable of swinging the operational axe as Oracle makes acquisitions to bolster its hardware business, and moves into services with gusto.
Hurd also knows about supply chains and wringing expenses out of them, although Oracle will likely never be a volume player like HP or Dell in the server racket. The need to simplify the supply chain and cut costs is probably why Oracle stopped making Sun Fire servers based on Opteron processors from Advanced Micro Devices and has cozied up to Intel. Sun might have focused on Opterons in the early 2000s for its initial x64 server push for much the same reason, but got burned by Opteron delays and was in the middle of switching allegiance to Intel when it ran upon the financial rocks.